On Wed, Apr 14, 2010 at 03:49:13PM +0300, Jaakko Heinonen wrote: > > Thank you for your comments. > > On 2010-04-14, Kostik Belousov wrote: > > The make_dev_credf() signature can be changed to return int error code > > instead of void, and a flag, lets call it MAKEDEV_CHECKNAME, added, > > make_dev_credf() returns struct cdev *, not void. Oops, forgot about this. We will need to change its signature more radically, but I still think that this worth it.
> > > that specifies that make_dev_credf() is allowed to fail with error > > EINVAL if supplied name is not sane. If flag is not supplied, > > make_dev_credf() should KASSERT or even just panic(9) if name > > does not pass internal validation. > > Otherwise sounds reasonable suggestion. However changes are needed in > devfs. Currently devfs_populate_loop() is not along the make_dev*() > code path but make_dev(9) increases a generation counter and > devfs_populate_loop() gets called sometimes later. To make things more > complex there may be several devfs mounts with different symlinks (files). Right, my point of my suggestion was to not modify populate_loop() at all. > > Actually my change doesn't work correctly in all cases with multiple > devfs mounts and has another bug with cdp use counts. I am considering > reverting the change. > > -- > Jaakko
pgpsso4iZPTm8.pgp
Description: PGP signature
